What is the maximum tolerance for a DIY Machine ?!

[Fancy]

I was wondering what have you seen achievable out there with DIY machines ? what tolerances were achieved and on what material ? I am asking to have realistic expectations of what is possible.

[Dave_stoke]

I was wondering what have you seen achievable out there with DIY machines ? what tolerances were achieved and on what material ? I am asking to have realistic expectations of what is possible.Can't answer for DIY but my China has a repeat cut tolerance of 0.05mm. Better when running a faster feed

Sent from my F8331 using Tapatalk

[hanermo2]

There are exactly zero limits.
Dan Gilabert made a 1 micron lathe.

My lathe has 1 micron mechanical resolution and positioning ability.
It is not natively accurate to 1 micron, but I can move by 1 micron to any offset size I can measure, using a digital 1 micron dti to confirm relative movement.
But I use thick ballscrews, 32 mm D, and 10.000 count ac brushless servos of 10Nm peak, 750W cont.
Lathes are about 10x more accurate, inherently, than milling.

This is about 800€ more for one axis than a typical stepper setup.
It is not too cheap, but not too expensive either, in context.

Anyone can make an xy machine with 0.01 mm repeatability, and 0.01 mm resolution.
A lot better is fairly easy to do by using
-bigger linear guides (stiffer, more rigid)
-thicker ballscrews (stiffer, more rigid)
-better screw mount (stiffer, more rigid) with better bearings.
-AC servos for more accuracy per rev

It is not that it is hard, as such.
But each improvement of those 4 costs about 100€ extra for one level and 200-250€ extra for 2 levels.
So for 2 levels or 2 sizes bigger/better you pay about 250€ x 4 == 1000€ extra.

So you can get 2 micron resolution for one axis, on a mill or router.
1 micron if you work more and do "better" in many ways.

So a 2 micron error, aka +/- 1 micron, is 2x2x2 = 8 microns volumetric local accuracy for milling/routing.
But screw errors tend to be about 0.1 mm, worst case, on typical 40-800 mm long axis travels.

So it is nowhere near inherently *accurate* to 0.01 mm (basic one axis), or 0.002 mm (very good axis), since the position is wrong.
But it can easily repeat to 2 microns, or 0.01 mm worst case, cheap.

Endless ways to contact probe, map, digital 1 micron dtis, glass scales, so you can offset the position to be accurate to 1 micron or less.
Granite surface plates, accurate to 2 microns, cost about 300€. 640x400 mm.
Gage blocks to make accurate probing points. Cheap to about 500 mm.
You can setup optical switches, accurate to about 1-2 microns, for less than 20€ each.

You can calibrate any screw, and map errors out to about 25x more inherent accuracy.
Thus you can offset the screw error in sw, by adjusting offsets or many ways in gcode, macros, probing, etc.
You can map the screw error in some controller sw, with various caveats.

It is not too hard or too expensive to map screws pretty well, more expensive is easier and more accurate.
Mapping to 0.01 mm positional error, per axis, is pretty easy.

Making stiffer screw mounts is easy, just make them 2 sizes bigger.

Example:
The ballnut mount on my lathe z axis is 120x120x70 mm tool steel. Yes, about 3" thick.
It was not *hard* to do, but took about 40 hours and 20€ in steel.
It is hard and slow to drill 120 mm deep, 10 mm, in tool steel.
12xD or 12 diameters deep.

Most accuracy is mostly making things bigger, thicker, more rigid, in steel.
After that some 30€ bearings, info, and work gets you better results.
E:
Like making a fixed-fixed screw, with tension, and a compliant tension mount.
This means you pull the screw about 500 kgf force, 1200 lbs, with an arrangement that allows the pulling end to move axially.
It cuts the free length of the screw by half.
This makes it 8x more rigid.
A tensioned screw is inherently 2x more rigid.

A 32 mm D screw == 1400 kgf max thrust, so pull about 500 kgf.

So you get 16x more rigidity for 60€ in materials and 2 days work.




I was wondering what have you seen achievable out there with DIY machines ? what tolerances were achieved and on what material ? I am asking to have realistic expectations of what is possible.

[Dave_stoke]

There are exactly zero limits.
Dan Gilabert made a 1 micron lathe.

My lathe has 1 micron mechanical resolution and positioning ability.
It is not natively accurate to 1 micron, but I can move by 1 micron to any offset size I can measure, using a digital 1 micron dti to confirm relative movement.
But I use thick ballscrews, 32 mm D, and 10.000 count ac brushless servos of 10Nm peak, 750W cont.
Lathes are about 10x more accurate, inherently, than milling.

This is about 800€ more for one axis than a typical stepper setup.
It is not too cheap, but not too expensive either, in context.

Anyone can make an xy machine with 0.01 mm repeatability, and 0.01 mm resolution.
A lot better is fairly easy to do by using
-bigger linear guides (stiffer, more rigid)
-thicker ballscrews (stiffer, more rigid)
-better screw mount (stiffer, more rigid) with better bearings.
-AC servos for more accuracy per rev

It is not that it is hard, as such.
But each improvement of those 4 costs about 100€ extra for one level and 200-250€ extra for 2 levels.
So for 2 levels or 2 sizes bigger/better you pay about 250€ x 4 == 1000€ extra.

So you can get 2 micron resolution for one axis, on a mill or router.
1 micron if you work more and do "better" in many ways.

So a 2 micron error, aka +/- 1 micron, is 2x2x2 = 8 microns volumetric local accuracy for milling/routing.
But screw errors tend to be about 0.1 mm, worst case, on typical 40-800 mm long axis travels.

So it is nowhere near inherently *accurate* to 0.01 mm (basic one axis), or 0.002 mm (very good axis), since the position is wrong.
But it can easily repeat to 2 microns, or 0.01 mm worst case, cheap.

Endless ways to contact probe, map, digital 1 micron dtis, glass scales, so you can offset the position to be accurate to 1 micron or less.
Granite surface plates, accurate to 2 microns, cost about 300€. 640x400 mm.
Gage blocks to make accurate probing points. Cheap to about 500 mm.
You can setup optical switches, accurate to about 1-2 microns, for less than 20€ each.

You can calibrate any screw, and map errors out to about 25x more inherent accuracy.
Thus you can offset the screw error in sw, by adjusting offsets or many ways in gcode, macros, probing, etc.
You can map the screw error in some controller sw, with various caveats.

It is not too hard or too expensive to map screws pretty well, more expensive is easier and more accurate.
Mapping to 0.01 mm positional error, per axis, is pretty easy.

Making stiffer screw mounts is easy, just make them 2 sizes bigger.

Example:
The ballnut mount on my lathe z axis is 120x120x70 mm tool steel. Yes, about 3" thick.
It was not *hard* to do, but took about 40 hours and 20€ in steel.
It is hard and slow to drill 120 mm deep, 10 mm, in tool steel.
12xD or 12 diameters deep.

Most accuracy is mostly making things bigger, thicker, more rigid, in steel.
After that some 30€ bearings, info, and work gets you better results.
E:
Like making a fixed-fixed screw, with tension, and a compliant tension mount.
This means you pull the screw about 500 kgf force, 1200 lbs, with an arrangement that allows the pulling end to move axially.
It cuts the free length of the screw by half.
This makes it 8x more rigid.
A tensioned screw is inherently 2x more rigid.

A 32 mm D screw == 1400 kgf max thrust, so pull about 500 kgf.

So you get 16x more rigidity for 60€ in materials and 2 days work.Best response to any questions I have seen in a long time.

Sent from my F8331 using Tapatalk

[Fancy]

Thank you both for replying to the thread. @hanermo2 you wrote a ton of stuff i am gonna need sometime to digest all that. could you elaborate with images preferably on this
Like making a fixed-fixed screw, with tension, and a compliant tension mount. ?? Thanks Guys.

[Robin Hewitt]

Let us consider one component, the run out on the collet chuck holding the tool... Standard precision 15-20 microns, Super precision 10 microns, Ultra precision 5 microns :culpability:

[Fancy]

That's sad to know :(

[routerdriver]

How many of us amateurs have workshops with sufficient temperature control to prevent thermal movement of both machine and workpiece becoming significant?For that matter,what kind of tolerance do most amateur projects require?It isn't so long since a DRO was an object of lust for most hobbyists.

[Neale]

Anyway, what kind of machine? Over what working area? While the figures quoted might be achievable on a milling machine with the rigidity that goes with it, put those ballscrews on a large working area router and they will hit performance and travel speeds, unless you pair them with massive motors. But that kind of machine is typically cutting wood, and there ain't much point in cutting that to micron accuracy.

So out of context, talking about achievable accuracy/resolution/repeatability and so on is interesting but not very useful. I think my 1500x750 cutting area router can theoretically achieve something like 3 micron resolution, based on step size and ballscrew pitch, but there is no way I would claim that it is accurate to that degree - and neither does it need to be for its intended purpose, cutting wood. Working to 3 thou, 25 times that theoretical resolution, is probably better than required. Good engineering is about trade-offs and compromise, not throwing money at the problem. Of course, a toolroom lathe or mill is a different question, with different criteria and different priorities - like accuracy over speed.

[Clive S]

Well said Neal. Timber with also change with moisture content from day to day. Plus the fact like a router built out of ali with change dimension as it warms up and cools down.

[Fancy]

I had the accuracy of a mill in mind of course not a large router.

[hanermo2]

Disagree.
All commercial collet chucks run out at less than 0.01 mm, way out from the spindle.

A typical iscar or similar collet chuck with tool mounted and tool, would run out maybe 5-8 microns 100 mm out from a typical ISO40/ISO30 spindle nose.

A kaiser, bigx, schaublin, regofix, or anything good typically promises about 2-4 microns 5-10 cm out from the spindle.
Lots of machines, tools, and setups do better.

Nothing commercial runs out over 10 microns or 0.01 mm because it would break, shatter, or destroy life of all modern carbide tooling and diamond tooling.
Standard basic TIR guarantee on machine tool spindles has been around 2 microns for a long time.

All high end expensive machines do much better, like Heller (German 5 axis), Fehlmann, or any japanese maker (Mori Seiki, DMG, etc).
Anyone can lap/fit/hone a tool and cone and taper to better than 0.01 mm / 5 cm length at home.
Anyone professional can do much better.
Amateurs hand-fit telescope lenses and mirrors to 30x better with no measurement equipment all the time.

Hand lapping with rigid laps gets easily better than 1 micron accuracy in size, since about 1940.

Gage blocks are machine lapped, cheap, and are lapped to 0.01 microns.
According to moore&wright, premier authority on the planet.

The runout on a tool mounted in a HAAS, ER collet chuck, needed to be 1-5 microns at 10.000 rpm trending low.
To engrave custom electrodes in carbon fiber for EGM.
I held the tool, the factory demonstrated success, I showed the tools and electrodes to press and about 200 industry reps. in us opening a 2011 year Barcelona HFO.
Best in the world HFO, sales, 2012.

Yes, I personally asked the factory to run the test and demo and we got back samples and video.
The tool was maybe 2 mm thick and 100 mm long and the tip was only 0.02 mm.
The tools and customer cases were in glass cases visible to several hundred visitors and pics with Gene Haas and US ambassador G. Philps and us are online in publications.

Professional ethics and courtesy and practice are why I don´t share pics.



Let us consider one component, the run out on the collet chuck holding the tool... Standard precision 15-20 microns, Super precision 10 microns, Ultra precision 5 microns :culpability:

[Colin Barron]

To achieve a decent level of accuracy heavy castings are handy to hold the tool or grinding wheel in position, preventing deflection and vibration. Even with grinding temperature control is not usually required, because the part used as the comparator can be kept at a similar temperature to the parts by immersion in coolant. The real accuracy trade off is space available and how much weight you can manage to install.

[Muzzer]

That's a wordy response (can't be bothered to quote it) that appears to make some sense on the face of it until you scratch a micron below the surface and twig that there are some rather big dimensions missing.

Unless you regulate the temperature of the table, frame and ballscrews, the coefficient of expansion becomes a significant issue. So if (like me) you have worked with suppliers that genuinely have 1um machines, you will see measures to address it. The machines have extensive heating and cooling circuits running continually and they are housed in air conditioned rooms. They are set up very painstakingly by experts who know what they are doing.

Similarly, a machine that won't deflect by more than a total of +/- 0.5um under load ("within 1um") would require a truly impressive level of rigidity. To demonstrate willy waving levels of accuracy outside of the pub, you need to machine a part under realistic conditions and measure the results - over several samples, ideally in several locations.

I doubt if any of us could even set up a machine to anything like the required level of accuracy before we even set up the work.

For an insight into what you need to get to the micron level, here's a mate of mine describing some of the issues. About half way in he's discussing 1um machines:

https://www.youtube.com/watch?v=9XRnVvJTj20

My big Japanese machine was originally designed and manufactured as a proper CNC machine with HSK bearings and ballscrews with a published achievable overall accuracy of 10um. Having said that, you'd need to allow for wear and adjust it carefully to get anywhere near that. I may achieve perhaps 25um on a good day and be happy enough with it. I'm not making watch or optical parts on it.

What could you possibly make that requires 1um tolerances? But above all - how would you measure them?

[JAZZCNC]

To demonstrate willy waving levels of accuracy outside of the pub,

I was having a shitty morning but that made me chuckle...Thank's.

Couldn't agree more, people are off in La La land when they quote levels of accuracy they can't even measure.! . . . . The floor they are standing on will move more than they quote as they walk around and it heats up.!

[m_c]

I miss JohnS in these kind of discussions.
I always remember a discussion somewhere (possibly on a ye olde usenet group), where somebody wanted to machine a bit steel to exactly 12" long, to within a stupid tolerance of a couple tenths.
John simply asked at what temperature did he want the tolerance met?
The guy didn't have a clue, and had to have it explained to him that for every degree of temperature change, the bar would grow/shrink by over a thou.

[Colin Barron]

To get these levels of accuracy, grinding with in process measuring is required, or make a batch and reject/rework the ones out of tolerance.

[RobC]

Holding 0.01mm most of the time, often can be on the measurement I wanted. But my machine is far from perfect, I've had to do a lot of tramming and shimming to achieve such, I built my machine using a drill press and a myford lathe which too isn't perfect. The goal is, if you can't make it perfect, make it adjustable!

This is my first build from no knowledge at all to being able to run it without a single problem which has taken me a few years.

28302

[Kitwn]

What a fabulous video! I want to be reborn, work harder at school and go to Cranfield University!

Half a um is the wavelength of blue light. Just measuring to that accuracy is serious science in it's own right and then he says they can make parts accurate to a few nm! One of my colleagues from a past life ('Dr Dave') had got his PhD by 'photographing' the individual atoms in crystal layers without an electron microscope. In effect he was measuring the varying distance between a microscopic metal probe and the surface of the atoms (if an atom can be said to have a surface) as the probe was moved over the crystal . He gave us a talk on how it was done. Nearly 30 years later I still remember sitting there utterly gobsmacked.

[JAZZCNC]

I miss JohnS in these kind of discussions.

Me too, I often see posts asking things like this and often think "I wish John was here to read this shit" . .. I can almost hear his sarcastic brutal reply's...:angel:

[alanambrose]

Yeah v interesting answer.

>>> You can setup optical switches, accurate to about 1-2 microns, for less than 20€ each.

Would love to know which ones?

TIA, Alan

[m_c]

Yeah v interesting answer.

>>> You can setup optical switches, accurate to about 1-2 microns, for less than 20€ each.

Would love to know which ones?

TIA, Alan

Probably the same ones that exist only in Hanermo's mind.
He loves posting all this on various forums, but has never yet produced a single bit evidence any of his machines exist anywhere other than in his own personal universe.

[Colin Barron]

Holding 0.01mm most of the time, often can be on the measurement I wanted. But my machine is far from perfect, I've had to do a lot of tramming and shimming to achieve such, I built my machine using a drill press and a myford lathe which too isn't perfect. The goal is, if you can't make it perfect, make it adjustable!

This is my first build from no knowledge at all to being able to run it without a single problem which has taken me a few years.

28302

Machining within 0.01mm (10 microns) sounds a little optimistic with those slideways and what about ballscrew accuracy and vibration?

[Muzzer]

As I said, the bottom line is actually putting a piece of steel etc on the machine and cutting the damned thing. A proper test piece will exercise the machine in all axes, with internal and external surfaces, circular and flat features etc.

Effects like backlash and machine flex happen when cutting real parts rather than on unloaded machines cutting air. Measure how the test piece compares against the model including circularity, flatness, perpendicularity, dimensional accuracy etc. It would be interesting to hear how the lunchtime legends got on....

[JAZZCNC]

Holding 0.01mm most of the time,

28302

Another one who's in La La land with Hanermo.! . . . Thou unlike him you have at least shown us your machine with its nice cheap Chinese round rails which tell us everything. If you are getting those readings then Suggest you get new measuring devices.

[MartinS]

Another one who's in La La land with Hanermo.!

Just to add some fuel....

We used to regularly machine to tolerances of less than 10 microns. But we had to spend £1m on a Swiss machine with a Granitan bed, BUT also had to spend a LOAD of £ on the measuring system and environment to prove it.

A little story:

Years ago we had a customer who produced(s) large diesel engines. They always had problems getting one bearing bore in the block correct. They were working towards TS16949 Automotive Quality Systems approval and someone new to their company who was bringing new systems in, looked at this issue. He discovered that the machining of the bore in the block wasn't the problem, it was the measuring system that was the cause of rejections! He changed the measuring system and the problem disappeared.

Our company was being pushed by this same company to introduce APQP (Advanced Product Quality Planning) systems. Much of what it contained it had good intentions but was "jobs for the boys" (cynical view), BUT a few core elements opened my eyes.

The CAPABILITY of a machine to produce a feature or shape

The CAPABILITY of a measuring system to "know" what a size or shape is.

The "door in the back of the wardrobe" moment:

So, I brought in "Gauge R & R" studies (Repeatability and Reproducibility). This resulted in us better trusting measurements.....some new equipment, fewer micrometers, even fewer verniers (1/10th mics not thou' mics etc), bore air gauges or Bowers in place of internal mics .

Then I had our guys perform "Machine Capability" studies. WOW what did we learn! They pointed us to many changes, some as simple as temperature control of coolant, or larger tanks. Different styles of cutters (tips or grinding wheels), more rigid fixturing, better maintenance of machines etc.

Eventually, our operators bonus earning increased, we re-deployed some of our inspectors (I renamed "Inspection Department" to "Quality Control"), our main customer relied on us feeding him our analysed quality data rather that his goods inwards inspection, and we made more profit. Yeeay win, win, win.

I know, a long story......

Many of us, in the past and now, "know" a machine, ie how much to "adjust", "tap", "nudge" it to get the job right (and part of me hopes that doesn't go away!)

But we have to understand what is ACTUALLY going on.

BTW, as I see it:

A machine does not have a "tolerance" when producing parts. It has a "capability" ie accuracy and precision.

The part has a tolerance which is wrapped around the maximum error that a feature can tolerate whilst still able to function correctly.


Like I said......just a bit of fuel (Dean).......:devilish::devilish::devilish:

[JAZZCNC]

Like I said......just a bit of fuel (Dean).......:devilish::devilish::devilish:

Martin since when have I needed any fueling. . . . I'm always gassed up with nitro and ready to blow....:whistle:

People just need to get real and keep the BS in their own little fantasy land. It's like Muzzer said the difference between moving around in fresh air and pushing through a material is light years different and most of those quoting micron tolerances have a machine which couldn't hold the claimed tolerances if pushing through butter..:encouragement: